of Identification Learning Disabilities in Spanishspeaking English Language Learners
David J. Francis University of Houston
Research supported by grants # P01HD39521 and R305U010001 from the National Institute of Child Health and Human Development and the Institute of Education Sciences.
Oracy/Literacy Development of Spanish-Speaking Children Biological and Behavioral Variation in the Language Development of Spanish-speaking Children
Coleen Carlson, University of Houston Jack M. Fletcher, University of Texas-Houston Barbara R. Foorman, University of Texas-Houston Claude N. Goldenberg, Cal. State U. - Long Beach Aquiles Iglesias, Temple University Jon Miller, University of Wisconsin Andrew Papanicolaou, University of Texas-Houston Panagiotis Simos, University of Texas-Houston Sharon Vaughn, University of Texas-Austin
Grants # P01HD39521 and R305U010001, National Institute of Child Health and Human Development and the Institute of Education Sciences.
Challenges in the Assessment of ELL Students
Incomplete overlap (i.e., holes) in the language and concept maps of children necessitate assessment in both languages Assessment in only one language will give an incomplete picture of students' knowledge, skills, abilities, and instructional needs Assessment in L2 can be hindered by students' failure to fully understand the instructions despite competence to perform the task in L2
Challenges in the Assessment of ELL Students
The availability of comparable assessments in L1 and in English is more the exception than the rule
Assesses the same domain of knowledge or skill Assesses the domain at comparable levels Assess the domain with comparable precision
Development of high quality L1 assessments is technically and financially demanding
Challenges in the Assessment of ELL Students
Comparable assessments may exist in L1, but individuals competent to conduct the assessment may not be available
Many urban school districts today report over 100 spoken languages among their students
Dialectical variation can contribute to the difficulties of building comparable tests and to finding suitable examiners
Challenges in the Identification of Learning Disabilities among ELL Students
Made more complex by limitations in the theoretical and empirical research on the development of academic competencies in ELLs
Support for common underlying models of literacy acquisition across alphabetic languages Unclear are
the degree of transfer that can be expected from L1 to L2, the conditions affecting transfer, and the effects of instruction and context on the strength of relationships in the model
Challenges in the Identification of Learning Disabilities among ELL Students
Clarifying these relationships also requires the development of comparable assessments across languages and explicitly crosslinguistic research
Domains where comparable L1 and L2 assessments are being developed
Letter names and sounds Phonemic Awareness Rapid Naming Untimed Word Reading Fluency - Word Reading Efficiency Comprehension
Light Green italic font indicates joint effort with Center for Applied Linguistics
Word Reading Untimed Graded Word Lists in Spanish
Chosen from K G3 instructional cumulative vocabulary materials Matched to database of word frequencies in print in Spanish texts (similar to CDR or Zeno, et al.) Selected words from K-3 with varying probability (low probability for K words; for non-K words, probability of selection varied by log(frequency), higher for low frequency words, lower for high frequency words
Fluency... Word Reading Efficiency Lists in Spanish
Selected from same corpus as untimed lists Multiple Formats and Multiple Forms of Each
Format A - Graded lists similar to TOWRE Format C - Non-graded list selected from K and G1 words
Students read for specified amount of time (45 seconds)
Connected Text (in development Tejas Lee)
Comprehension
Goal: assess comprehension processes as independently as possible Strategy:
keep decoding demands minimal make the purpose of the reading activity clear reduce the effect of variation in vocabulary and background knowledge
Comprehension
Yields measures of four component processes
Text Memory Access to Background Knowledge Text Inference Knowledge Integration (inference
Phonological Processes in Spanish: TOPP-S
Phonological Awareness
Sound Matching (Initial Sounds / Final Sounds) Blending Phonemes into Words Blending Phonemes into Non-words Phoneme Elision Segmenting Phonemes into Words Segmenting Phonemes into Non-words
Phonological Memory
Non-word Repetition Digits Reversed
RAN Letters and Numbers
Subtest Design
Each subtest consists of multiple items (10-20) and separate practice items With the exception of First and Final Sound Matching, all subtests are built entirely of production based items Items are targeted to match CTOPP items linguistic in complexity (# of phonemes; area of manipulation) but be appropriate for Spanish
Table of Sample Sizes and Reliability Estimates for Seven PA Subtests (N=1,705)
Coefficient Alpha Sample Size (standardized) Measure First Sound Matching Final Sound Matching Blend Phonemes intoWords Blend Phonemes into Non-words Phoneme Elision Segmenting Words into Phonemes Segmenting Non-Words into Phonemes Total Test Spanish 1,116 1,069 1,435 1,123 1,178 1,227 983 1,477 English 965 902 1,045 985 893 967 917 1,076 Spanish 0.97 0.94 0.94 0.93 0.97 0.96 0.96 0.99 English 0.98 0.95 0.96 0.93 0.96 0.95 0.95 0.99
English SUBTEST Oral Language Oral Language Reading Comp English Phonemic Awareness Word Reading Word Reading Efficiency Oral Language Reading Comp Spanish Phonemic Awareness Word Reading Word Reading Efficiency 1.00 Reading Comp 0.79 Phonemic Awareness 0.72 Word Reading 0.74 Word Reading Efficiency 0.65 Oral Language 0.16 Reading Comp 0.22
Spanish Phonemic Awareness 0.40 Word Reading 0.12 Word Reading Efficiency 0.12
1.00
0.69
0.82
0.78
0.26
0.39
0.46
0.22
0.27
1.00
0.76
0.66
0.31
0.38
0.65
0.36
0.24
1.00
0.88
0.38
0.51
0.64
0.51
0.47
1.00
0.34
0.49
0.54
0.43
0.52
1.00
0.78
0.64
0.71
0.67
1.00
0.70
0.83
0.81
1.00
0.72
0.58
1.00
0.81
1.00
Bilingual Language & Literacy Project Language Sample Data Variables
Variable NCW Language Measure No. Complete Words Description NCW provides an index of transcript length in terms of words. All words are counted, including words in mazes (filled pauses, false starts, repetitions, and reformulations) and words produced in the non-target language. WPM is a measure of speaking rate and has been used as a measure of language proficiency. The more proficient the child is in the language, the higher the WPM. NTW counts the total number of words that contribute to telling the story. Words in mazes and parenthetical remarks are not included in this count. Neither are words in incomplete or unintelligible utterances. NTW is broken down into English words (NTWE) and Spanish words (NTWS) for each language sample. NDW counts the different words within the set of total words (NTW) and is broken down into English words (NDWE) and Spanish words (NDWS). NDW is a wellestablished measure of vocabulary diversity and is highly associated with advancing age. MLUW measures the length of the utterances in words. The more proficient the child is, the longer the utterances s/he produces. This is a general measure of syntactic complexity and correlates very highly with age. RUBRIC is a narrative structure analysis consisting of seven components: introduction, character development, mental states, referencing, conflict resolution, cohesion and conclusion. Each component is given a score (0-5). RUBRIC is the total score (0-35). The higher the score, the better the story. SI is a measure of the syntactic complexity of the child's utterances, calculating the ratio of number of clauses to number of utterances. The higher the SI value, the more subordinating clauses used. 1 Notes
WPM NTW NTWE NTWS NDW NDWE NDWS MLUW
Words/Minute No. Total Words No. Total Words-English No. Total Word-Spanish No. Different Words No. Diff. Words-English No. Diff. Words-Spanish MLU in Words
1, 2
1, 2
1
RUBRIC
Story Structure Rubric
SI
Subordination Index
1. Calculations are based only on the complete and intelligible utterances. Utterances that contain unintelligible segments or have been abandoned or interrupted are not included. 2. A high number for NDW and NTW in the non-target language is an indicator of deficiency in the target language, e.g., as many as 30% of the words used to tell the story in English are Spanish words and this number decreases with advancing grade. The reverse, however, is not necessarily true. A low number for NDW and NTW in the non-target language may not indicate a proficiency in the target language as some children do not choose to code switch.
Subjects with English and Spanish Narratives in Retell and Unique Conditions
LANGUAGE ENGLISH ENGLISH ENGLISH ENGLISH SPANISH SPANISH SPANISH SPANISH GRADE 1 2 3 K 1 2 3 K Retell
174 192 193 121 181 165 166 153
Unique
168 187 192 115 176 167 173 150
Comparison of English and Spanish on four Standardized Assessments
r = -.336
r = .305
r = -.084
r = -.135
Comparison of English and Spanish Narratives - Retell
r = .392
r = .437
r = .410
r = .513
Comparison of English and Spanish Narratives - Unique
r = .447
r = .499
r = .440
r = .537
Conclusions
Identification of Learning Disabilities in ELLs requires understanding of linkages among theoretical variables of interest The moderating effects of instruction and context on these linkages Development of comparable assessments in multiple languages Knowledge of the operational characteristics of key measures used in identification